Geofoam Applications in the Design and Construction of Highway Embankments (2004)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

4-1 CHAPTER 4 PAVEMENT SYSTEM DESIGN Contents Introduction ................................................................................................................................................. 4-1 Benefits of a Thicker Pavement System...................................................................................................... 4-3 Increase in the Life of the Pavement ....................................................................................................... 4-3 Increase in Internal Stability.................................................................................................................... 4-4 Differential Icing ..................................................................................................................................... 4-5 Solar Heating........................................................................................................................................... 4-6 Accommodation of Utilities and Road Hardware ................................................................................... 4-7 Drawbacks of a Thicker Pavement System ................................................................................................. 4-7 Utilizing Geofoam Layers with Varying Properties.................................................................................... 4-7 Separation Materials.................................................................................................................................... 4-8 Pavement Design Procedures .................................................................................................................... 4-16 Flexible Pavement System Design Catalog............................................................................................... 4-20 Rigid Pavement System Design Catalog ................................................................................................... 4-24 Summary of Pavement System Data From Case Histories........................................................................ 4-26 Typical Dead Load Stress Range Imposed By A Pavement System......................................................... 4-27 Summary ................................................................................................................................................... 4-30 References ................................................................................................................................................. 4-31 Figures....................................................................................................................................................... 4-34 Tables ........................................................................................................................................................ 4-36 ______________________________________________________________________________ INTRODUCTION The objective of pavement system design is to select the most economical arrangement and thickness of pavement materials for the subgrade provided by the underlying expanded

4-2 polystyrene (EPS) blocks. The design criteria are to prevent premature failure of the pavement system, as defined by rutting, cracking, or similar criterion, which is a SLS type of failure, as well as to minimize the potential for differential icing (a potential safety hazard) in those areas where climatic conditions make this a potential problem. Also, when designing the pavement cross- section overall, consideration must be given to providing sufficient support, either by direct embedment or structural anchorage, for any road hardware (guardrails, barriers, median dividers, lighting, signage and utilities). However, a unique aspect of pavement design over lightweight fill is that the design must also consider the external and internal stability of the embankment. The pavement system is defined as including all materials, bound and unbound, placed above the EPS blocks. Table 4.1 provides a summary of EPS-block geofoam engineering parameters typically used in pavement design. Resilient modulus values for typical EPS densities range from 5 to 10 MPa (725 to 1,450 lbs/in²) and the corresponding California Bearing Ratio (CBR) values range between 2 and 4 percent. Therefore, depending on the design traffic loads and desired pavement life a relatively thick pavement system, minimum 610 mm (24 in.), will be required for embankments containing EPS-block geofoam. The benefits of using a thicker pavement system include increased pavement life, increased internal stability of the embankment (see Table 3.1), reduced potential for differential icing, reduced potential for solar heating, and better accommodation of shallow utilities and road hardware. The drawbacks of a thicker pavement system include increased weight which will decrease external stability of the embankment (see Table 3.1). Thus, some compromise is required to optimize the final design of both the pavement system and overall fill. The benefits and drawbacks of utilizing a thicker pavement system are further discussed in the subsequent sections. Procedures for design of pavement systems over EPS-block geofoam embankments are also summarized and recommendations regarding

4-3 preliminary pavement system thickness and material unit weights to utilize for preliminary external and internal stability analysis are provided. Table 4.1. Equivalent Soil Subgrade Values of EPS-Block Geofoam for Pavement Design. This chapter presents detailed information on the pavement design aspect of the EPS- block geofoam design methodology. An abbreviated form of the pavement design procedure can be found in the provisional design guideline included in Appendix B. BENEFITS OF A THICKER PAVEMENT SYSTEM Increase in the Life of the Pavement The objective of “flexible” pavement design using asphaltic-cement concrete (AC) is to determine the type, thickness, and arrangement of the various pavement layers to minimize critical pavement responses and provide a serviceable pavement for the intended design life (1). Pavement responses are the stresses, strains, and deflections that occur in AC pavements from traffic loading, daily or seasonal temperature variations, moisture changes, and changes in the pavement support conditions. Critical pavement responses are those material responses which, through a single or repeated occurrence, will result in structural deterioration of the pavement (1). Critical AC pavement responses include the tensile strain at the bottom of the AC or stabilized base layer, the vertical stress on the top of any unbound granular layer, and the vertical stress at the top of the subgrade (1). Each layer of the pavement system must be designed to prevent overstressing of the underlying layers. The load that a pavement layer can carry is related to the stiffness of the layer. Stiffness of a material is its ability to resist deformation within the linear elastic range of the stress-strain relationship (2). It is the magnitude of force that must be applied at some point to produce a unit displacement at that point (3). Stiffness of a pavement layer is often measured by E∗D3 where E is the modulus of elasticity of the layer and D is the thickness of the layer (1). This relationship indicates that the thickness has more influence than the material modulus in the performance of the layer. Thus, the best method of preventing overstressing of the

4-4 underlying layers is primarily accomplished by increasing the thickness of selected upper layer(s). A thicker layer will also reduce the tensile stresses and strains at the bottom of a stiff layer that is placed on a less stiff layer as is typically the case in pavement sections. Additionally, the most cost effective method of reducing the compressive stresses in the subgrade is to increase the thickness of the subbase or the layer that provides the most stiffness for the least cost (1). The objective of “rigid” pavement design using portland-cement concrete (PCC) is to determine the slab thickness, base type, joint spacing, load transfer, and drainage that will limit stresses in the PCC pavement and that will result in the lowest annual cost, as shown by both initial construction costs and future maintenance costs. Stresses in PCC slabs result from traffic loads and from environmental sources. Environmental sources of stress include thermal gradients, moisture gradients, drying shrinkage, thermal expansion and contraction, and foundation movements. Critical PCC pavement responses include the maximum bending stress due to load and curl along the longitudinal shoulder joint, midway between transverse joints; the combined load and curl stress along the transverse joints; the maximum corner deflection; and the maximum bending stress between the concrete and dowels (1). Since PCC is stronger in compression than tension, tensile stresses that develop from these responses are of primary concern. Stresses in the slab can be reduced by increasing the thickness of the PCC slab and/or subbase. Additionally, shorter joint spacing and drainage improvements can also reduce stresses in the slab. Increase in Internal Stability The thickness of the pavement system affects the internal stability of the embankment by two mechanisms. As the thickness of the pavement system increases, the dead-load stress on the EPS increases. However, as the thickness of the pavement increases and the corresponding distance between the pavement surface and the top of the EPS mass increases, the live-load stress on the EPS from vehicle tires decreases. Theoretically, there is an optimum pavement thickness for which the combined dead- and live-load stresses are minimized. This optimum pavement

4-5 thickness depends on the unit weights of the pavement materials and the pressure magnitude and footprint area of the vehicle tire. Differential Icing Lightweight fill materials, especially non-earth type of materials, have thermal properties that differ from earth materials. The difference in thermal properties will result in the pavement surface overlying lightweight fills to have a temperature different than the pavement surface overlying a soil subgrade. The non-earth pavement section will generally be cooler in winter and warmer in summer (4). Two consequences of the variance in thermal properties include the development of differential icing conditions during cold temperatures and increased solar heating of the pavement surface during warm temperatures. The potential safety and maintenance costs associated with the thermal properties of non-earth materials is rarely considered in practice at the present time (4). Differential icing is caused by the formation of ice on the pavement surface that is underlain by non-earth material adjacent to pavement underlain by soil that is ice free (4). Differential icing is similar to the “bridge deck” problem whereby the bridge deck freezes before the adjacent road sections in that they both are considered safety hazards (5). Thus, differential icing must be considered in the design of the pavement system. A detailed discussion on differential icing, which is included in (5), indicates that based on the detailed studies in Norway (6) and Sweden (7), differential icing can be minimized by providing a sufficient thickness of soil between the top of the EPS mass and the top of the pavement surface. Differential icing can also be minimized by utilizing base and/or subbase materials with sufficient “fines” to hold water, which has a relatively high heat capacity, to provide sufficient retained heat to keep the pavement surface as close to the temperature of the adjacent pavement sections. Fine materials are materials with particle size less than the No. 200 sieve. However, specific recommendations for material particle distributions to include the minimum amount of fines is currently unavailable.

4-6 Because using base and/or subbase materials with fines affects the structural quality of the base and/or subbase and due to the lack of standardized tests to determine thermal properties of non- earth materials as well as the lack of a thermal design procedure that considers pavement material thickness and composition, it is recommended that a sufficient thickness of soil between the top of the EPS mass and the top of the pavement surface be used. Based on the minimum recommended pavement system thickness from the Norwegian design guidelines (5,8) of 400 mm (16 in.) to 800 mm (32 in.) and the Swedish guidelines of 400 mm (16 in.) to 500 mm (20 in.) (5,7), it is recommended that a minimum pavement system thickness of 610 mm (24 in.) be used with EPS blocks. It should be noted that the type of separation layer if any, used between the EPS and pavement system might affect the potential for differential icing. In particular, the use of a PCC slab as a separation layer may actually increase the potential for differential icing. Solar Heating The pavement surface overlying lightweight fills will generally be warmer in the summer than a pavement surface overlying a soil subgrade (4). Solar heating occurs due to solar heat becoming trapped within the pavement section due to the thermal-insulative characteristics of non-earth lightweight materials (4). It is anticipated that solar heating of AC pavements will be greater than that of PCC pavements. This solar heating may accelerate deterioration of the AC pavement layer(s) due to the decrease in the Young’s modulus of the AC layer with an increase in temperature. As the modulus decreases, the tensile strain that occurs within the AC with each application of a vehicle load increases. Fatigue failure of AC is linked to the magnitude of tensile strain within the AC under each load application, i.e., the more strain per load cycle the fewer load cycles to achieve a given level of permanent pavement distortion (cracking or rutting). Although there has not been any explicit study of whether or not AC pavements underlain by EPS-block geofoam experience premature failure, based on the differential icing studies, it is

4-7 anticipated that the greater the distance between the top of the pavement and top of the EPS mass, the less of an increase in pavement surface temperature. Accommodation of Utilities and Road Hardware Shallow utilities and road hardware (barriers and dividers, light poles, signage) can be accommodated by providing a sufficient thickness of the pavement system to allow conventional burial or embedment within soil or, in the case of appurtenant elements, by anchoring to a PCC slab or footing that is constructed within the pavement section. DRAWBACKS OF A THICKER PAVEMENT SYSTEM A thicker pavement system will yield a heavier pavement system. This increase in weight will affect the external stability of the embankment through increased settlement and reduced stability. It is possible to reduce the total thickness of the pavement system and still provide a technically efficient pavement system by utilizing an EPS with a higher density for the uppermost layer of blocks within the horizontal limits of the paved area and by providing a separation material between the EPS and pavement system that will provide reinforcement of an unbound pavement layer placed directly above the separation material. The subsequent paragraphs discuss these two alternatives for decreasing the thickness of the pavement system. UTILIZING GEOFOAM LAYERS WITH VARYING PROPERTIES Two possible reasons for utilizing geofoam layers with varying properties in embankment construction are provided below. • allow a minimum pavement section thickness by utilizing an EPS with a higher density, e.g., EPS100, for the uppermost layer of blocks within the horizontal limits of the paved area. The resilient modulus of the EPS blocks directly underlying the pavement system influences the thickness of the proposed pavement system. The use of an EPS with a higher density, and thus higher resilient modulus, in the upper most layer of blocks will allow a thinner

4-8 pavement system thickness. However, the predominant practice in the U.S.A. is to utilize a single EPS density for the whole embankment. An investigation into the benefits of using a higher density geofoam layer in the uppermost layer of blocks is required so that a cost-benefit analysis of providing a higher density geofoam versus providing a thicker pavement section can be performed. Preliminary indications are that although the use of a higher density EPS for the uppermost layer will yield a thinner pavement system, the thinner pavement section may not offset the higher cost of using a higher density EPS as shown in the cost data in Chapter 12 where the material cost increases at 55 percent for increasing density from EPS50 (lowest, recommended directly below paved areas) to EPS100 (highest). The impact of using a higher density EPS directly below the pavement system is further discussed later in this chapter. • allow a more cost effective embankment by using lower density blocks at greater depths. Because the total calculated vertical stresses will decrease with depth within the EPS mass as well as be less under the side slopes as opposed to beneath the paved area, it is possible to use multiple densities of EPS blocks, e.g. lower density blocks at greater depths and/or under side slopes. This will reduce overall costs of the EPS-block geofoam. However, for constructability it is recommended that no more than two different density EPS blocks be used on the same project. This use of different densities of EPS blocks affects internal stability and is further discussed in Chapter 6. SEPARATION MATERIALS A separation layer between the top of the EPS blocks and the overlying pavement system can have two functions.

4-9 • To enhance the overall performance and life of the pavement system by providing reinforcement, separation, and/or filtration. A separation layer used for these purposes is, technically, part of the pavement system. • To enhance the durability of the EPS blocks both during and after construction. A separation layer provides reinforcement to the unbound pavement layer by providing additional horizontal confinement. This additional confinement increases the strength and stiffness of the unbound layer because the stress-strain behavior of typical unbound layers, which typically consist of coarse-grained soil, is dependent on the effective confining stress. It is concluded in (9) that unbound roadbase materials have a lower "effective" stiffness in pavement structures without a cement-bound load-spreading layer above the EPS subbase compared to a pavement structure with a cement-treated capping layer. It is recommended in (10) that a modulus of elasticity value for the unbound material of up to 50 percent lower be assumed in linear-elastic design calculations if the unbound material is placed directly over an EPS subbase than typically assumed for unbound roadbase materials. A cement-treated capping layer is recommended for heavy-duty roads because it neutralizes the effects of open joints between the EPS blocks, provides support to overlying unbound base material under high traffic intensity, and eliminates any restriction for use of less-expensive, low density EPS types (10). The use of a 100 to 150 mm (4 to 6 in.) thick reinforced PCC slab is currently the state of practice primarily because it is considered to be a necessity for providing sufficient lateral confinement of unbound pavement layers when using EPS-block geofoam and because of historical usage of PCC slabs dating back to the earliest EPS-block geofoam lightweight fills in Norway in the 1970s. As a rule of thumb, the slab was found equivalent to unbound material on a 1-to-3 basis, i.e., 1 mm (0.04 in.) of slab replaced 3 mm (0.12 in.) of unbound pavement material although this equivalency has been disputed recently in (9). The original function of the PCC slab was for pavement reinforcement and the intent was to allow the use of a minimum pavement

4-10 system thickness. In later designs, the PCC slab was also used for the function of a barrier against potential petroleum spills. However, the use of a PCC slab for this function is questionable due the long-term development of cracks in PCC slabs. Problems with the use of a PCC slab include • The potential for sliding of the slab during an earthquake (the Japanese require L- shaped reinforcing bar dowels cast into the slab that penetrate down into the EPS blocks). • The potential for ponding of water within the pavement system. • The increased potential for differential icing and solar heating (due to both a thinner pavement system as well as the thermal properties of PCC). Additionally, PCC slabs generally represent a significant relative cost. It may be the only PCC work on a project (a major consideration in remote areas). The results of a cost analysis, which is included in Chapter 12, reveal that the cost of a reinforced slab is high and can range from 20 to 30 percent of the total project cost. A survey conducted as part of this study (see Appendix A) listed the cost of the PCC slab as one of the major cost concerns of U.S. engineers in current practice. Alternative separation layers for reinforcement that can be considered in pavement design include a geogrid, a reinforced geomembrane that will also resist hydro-carbon spills, geocell with soil or PCC fill, and soil cement. One application where a PCC slab is typically required is when an embankment with vertical sides, i.e., geofoam wall, is used (11). The primary function of the PCC slab is to support the upper part of the exterior facing system. A secondary function is to provide anchorage for various highway hardware such as safety barriers, signage, and lighting. A PCC slab used for these functions will act primarily as a structural member for the benefit of other embankment system components and not the EPS. Therefore, the PCC slab should be designed for the intended function.

4-11 The following is an assessment of the various separation layer alternatives that have either been used or proposed for use over the approximately 30 years that EPS-block geofoam has been used as lightweight fill for road embankments. They are listed in an approximate order of increasing complexity of construction (and, therefore, probable increasing cost): • No separation material. This is acceptable practice in a number of published design manuals in other countries (12-14) and therefore is the baseline against which all other alternatives are compared. • Separation Geotextile. This offers positive protection against soil particles from the pavement system migrating downward into any gaps between the EPS blocks. A geotextile is known to have been used on a least one road project in Germany where the lowest pavement layer was a fine sand. Interface friction angles between geotextiles and EPS block are provided in Chapter 2. • Geomembrane. This offers positive protection against both petroleum spills (assuming a petroleum-compatible polymer is chosen for the geomembrane) and soil particle migration. Interface friction angles between geomembranes and EPS block are provided in Chapter 2. For seismic considerations, it is proposed that a Newmark (15) sliding block analysis be conducted to assess the magnitude of seismically induced permanent deformation that could occur along the EPS/geomembrane interface. This analysis relates the interface friction to the magnitude of permanent deformation. The magnitude of permanent deformation should not exceed about 50 mm (2 in.). If it does, a different type of geomembrane should be used to increase the frictional resistance. Of course, the geomembrane should be installed and welded by a geomembrane installer that has experience with geomembranes, e.g., landfill liner experience, to reduce the

4-12 potential for geomembrane holes. However, the geomembrane surface should not be horizontal because liquid (including surface water infiltrating through the pavement system) could become trapped which could lead to strength loss of the pavement system due to pore-pressure buildup. It is of interest to note that the practice in the U.K. (14) is to put a 50 mm (2 in.) (minimum) thick sand bed directly on top of the EPS blocks and then put the geomembrane on the sand bed. The sand bed is crowned to allow gravity flow of any liquid to the sides of the fill. • Geogrid. This would only stiffen the unbound layer(s) of the pavement system. This alternative has not been researched to date (although the benefit for soil subgrades with CBRs similar to that of EPS-block geofoam has been studied) and there has been no known use of this to date. The geogrid would probably be placed not on top of the EPS blocks but within the lower half of the unbound layer(s) as research to date indicates that this is the most effective location. • Geocell with soil fill. This would have a primary benefit of stiffening the unbound layer(s) of the pavement system as well as providing a reinforced working surface as a constructability aid. This alternative has not been researched to date for EPS-block geofoam and there has been no known use of this to date. The geocell would be placed directly on top of the EPS blocks. • Soil cement. This technique involves mixing a relatively small percentage of portland cement with a coarse-grain soil to form what is basically a weak concrete. This would have a primary benefit of stiffening the unbound layer(s) of the pavement system as well as providing both a reinforced working surface as a constructability aid and some protection against petroleum spills (although ponding of water within the pavement system would be a concern as it is with a

4-13 geomembrane). An additional concern would be the potential for exacerbating differential icing due to the fact that heat is removed easily from concrete due to the minimal water content of the cement. This alternative has received limited research in Germany solely for its efficacy of stiffening the pavement system and enhancing pavement life. It is not known to have been used for actual fills to date. • Pozzolanic stabilized materials: These materials consist of flyash, an activator (lime, cement, lime-kiln dust, and/or cement-kiln-dust), and aggregate. The pros and cons of this alternative would be similar to those of soil cement. This alternative is not known to have been used for EPS fills to date. • Geocell with PCC fill. The pros and cons of this alternative would be similar to those of the soil cement. This alternative is not known to have been used for EPS fills to date. • Reinforced PCC slab. Although the alternative of no separation layer is the baseline case for technical and cost comparisons, use of a 100 to 150 mm (4 to 6 in.) thick reinforced PCC slab (which is at the opposite end of the design spectrum in terms of cost and complexity) is the state of practice even though mistakenly considered to be a necessity when using EPS-block geofoam. This is due to historical usage of PCC slabs dating back to the earliest EPS-block geofoam lightweight fills in Norway in the 1970s. However, contrary to a widely held belief, the reason for using a slab in Norway was originally and primarily for its pavement reinforcement benefit. As a result, it allowed a minimum pavement system thickness but this is now recognized as increasing the potential for differential icing. It was only later that its benefit to act as a barrier against potential petroleum spills was suggested. However, some have questioned this as

4-14 the long-term development of cracks in PCC slabs is not uncommon. Additional problems with using a PCC slab include sliding of the slab during an earthquake (the Japanese require L-shaped reinforcing bar dowels cast into the slab and penetrating down into the EPS blocks); ponding of water within the pavement system; and increased potential for differential icing (due to both a thinner pavement system as well as the thermal properties of PCC). The cost and the fact that it often requires the only PCC work on a project (a major consideration in remote areas) are construction concerns. In summary, a PCC slab is not required and if the cost is excessive one of the alternatives listed above should be used. In addition, if the roadway is not heavily loaded, the unbound pavement layer may not need additional confinement and thus only a geomembrane can be used to protect against hydrocarbon spills. The need for stiffening the unbound pavement layer(s) should be investigated on a project-specific basis to evaluate whether stiffening or a thicker unbound layer(s) is more cost effective. Some of these alternatives such as the use of a geocell will also provide a stiff working platform during construction and provide some protection of the EPS blocks during construction. The use of a soil-filled geocell is particularly promising because • It can be placed with relative ease and speed • It makes use of the granular soils that would be used anyway as part of the pavement system to fill within its cells. • It enhances constructability as it provides a working surface for placement of subsequent pavement layers. Migration of the finer soil particles from the unbound layer(s) of the pavement system into the gaps between the EPS blocks may occur due to gravity and erosion due to infiltration (rain or melted snow/ice runoff). Both migration and infiltration could lead to void formation and

4-15 settlement of the pavement system. A separation layer to prevent migration and filtration of the finer soil particles from the unbound layer(s) of the pavement system into the small gaps between EPS blocks is only required if the unbound pavement layer(s) contain a relatively large proportion of smaller particles (fine sand and smaller). Small gaps between blocks will occur due to manufacturing tolerances of the EPS blocks, human error, and sloppiness in block placement at the project site. Experience indicates that gaps of the order of 25 mm (1 in.) are acceptable in terms of performance of the overall fill. A geosynthetic separation layer could be used between the top of the EPS blocks and bottom of the unbound pavement layer to provide separation and filtration of the fine soil particles. An appropriate geotextile would be able to serve both of these functions. The primary durability concern for EPS in road embankments is the fact that liquid petroleum hydrocarbons (gasoline, diesel fuel/heating oil) will dissolve EPS if the EPS is inundated with the liquid. Therefore, the concern about the potential for a motor vehicle accident involving some type of fuel transport truck wherein large volumes of liquid petroleum hydrocarbons are released, seep downward into the fill, and dissolve the EPS blocks has been expressed in projects that have utilized or considered utilizing EPS. It is suggested that the potential for such an accident is relatively small. When such an accident occurs, it is likely that most of the petroleum is consumed in the fire that often ensues. The petroleum that does seep onto the ground would attack the uppermost pavement layer first (if it is AC as is the more common case). Further, any petroleum that seeped into the embankment would generally trigger an environmental concern requiring excavation and removal of any contaminated material. Therefore, concern over protecting EPS-block geofoam from petroleum spills does not appear great nor cost effective. This position is supported by the German national design manual (12,13) which is the most recent developed outside the U.S.A. Consideration may be given to performing

4-16 a risk analysis by obtaining petroleum spill occurrence data from a transportation agency or the Environmental Protection Agency (EPA). If protection of the EPS from vehicle fuel spills is desired, a geomembrane of appropriate composition can be used (14). For example, the specifications for the I-15 reconstruction specified a minimum 0.7 mm (28 mil) gasoline-resistant geomembrane manufactured from a tripolymer consisting of polyvinyl chloride, ethylene interpolymer alloy, and polyurethane or a comparable polymer combination. In summary a separation material between the EPS blocks and pavement system should not be used without a project-specific needs assessment due to the costs involved and the technical viability of not using any separation material. This needs assessment should be based primarily on two considerations: • The need to prevent soil particles from migrating downward between any small gaps between the EPS blocks. This will depend solely on the gradation of the unbound pavement layer directly on top of the EPS blocks and the care taken during construction to minimize the development of gaps along the vertical joints between EPS blocks. • The desire to stiffen the unbound pavement layer(s). This becomes more important as the overall thickness of the pavement system decreases and the traffic loads increase. PAVEMENT DESIGN PROCEDURES Traditional pavement design procedures may be used by considering the EPS to be an equivalent soil subgrade. The resilient modulus or equivalent CBR value of the EPS can be used in the design procedure. A summary of these design parameters is provided in Table 4.1. These CBR values, which range from 2 to 5 percent, were measured as part of the A47 Great Yarmouth Western Bypass project in the United Kingdom (14).

4-17 Results of field studies of flexible pavement systems over EPS-block geofoam are available in (9,16-20). The findings of these studies are incorporated in the current Dutch design manual for lightweight pavements with expanded polystyrene geofoam (10). A summary of these findings is provided below. • The horizontal strain at the bottom of an asphalt layer, is approximately 15 percent higher when an EPS subbase is used below an unbound roadbase compared to a sand subbase. This strain difference results in a two times lower allowable number of standard axle load repetitions or a two times shorter pavement design life (9). • Because of the low modulus of elasticity of EPS, the thickness of the EPS has only a marginal influence on the pavement behavior under loading and on the pavement design life. The stress and strain values in the upper pavement layers are similar in pavement structures with different thicknesses of the EPS layer (9). • The presence of an EPS subbase in a pavement structure has a significant influence on stress and strain development in the pavement. Unbound roadbase materials have a lower ‘effective’ stiffness in pavement structures without a cement-bound separation layer above the EPS subbase compared to the stiffness of unbound materials placed over a cement-based separation layer. Although not specifically indicated in (10), the most likely reason for these differences in stiffness is that better compaction of the unbound material can be achieved over a cement-bound separation layer than directly over EPS blocks. Therefore, in linear-elastic calculations the assumed modulus of elasticity value for the unbound material should be lower when this material is laid directly over an EPS subbase than when the material is placed over a cement-bound layer (9). It is

4-18 recommended that the modulus of elasticity of unbound materials placed directly over EPS blocks be reduced by up to 50 percent in design. (10). • Implementation of a cement-treated separation layer on top of the EPS subbase substantially increases the design life of the pavement structure with an EPS subbase. The application of such a layer is therefore recommended for pavement structures subjected to heavy traffic loading (9). As noted earlier, this does not mean that a reinforced PCC slab is required in all applications. Results of field studies of rigid pavement systems over EPS-block geofoam could not be found in the current literature probably because EPS-block geofoam is typically utilized over soft soils where flexible pavements are traditionally used. However, the Interstate 15 reconstruction project in Salt Lake City, Utah used a rigid pavement system over EPS blocks. Therefore, long- term rigid pavement performance will be available in the near future. Two approaches to pavement system design that can be used for the design of pavements over EPS-block geofoam include mechanistic-empirical methods and empirical-regression methods. Mechanistic analyses involve the external calculation of stresses and strains in various critical regions of a pavement cross-section due to external application of a load, i.e., traffic, using appropriate modeling tools such as an elastic layer analysis or finite element analysis (21). The calculated stresses and strains are compared with values known from experimental or theoretical studies to be the maximum allowable, based on predictions of pavement performance (physical distress, such as cracking, rutting, or roughness) (1). The pavement can then be designed by adjusting the different layer thicknesses so that the calculated responses are less than the allowable maximum values. The Shell pavement design procedure (22) is an example of a mechanistic-empirical procedure. The AASHTO 1993 pavement design procedure (23) is an example of an empirical-regression method. The AASHTO pavement design procedure is based on data obtained from a series of road tests and equations obtained from regression methods.

4-19 However, it is anticipated that the revised AASHTO 1993 design procedure will include mechanistic-empirical procedures. The Dutch pavement design procedure is based on the Shell Pavement design procedure (22), which is a mechanistic procedure that is based on limiting the horizontal tensile strains at the bottom of the asphalt layer in order to prevent asphalt fatigue cracking and limiting vertical compressive strains at the top of the soil subgrade to prevent excessive permanent deformation in the subgrade. However, the Shell Pavement design procedure was modified to include an EPS subbase. The Dutch pavement design procedure is based on limiting elastic deformation in the EPS subbase due to cyclic (traffic) loading to 0.4 percent (10). The Dutch design procedure consists of explicitly calculating the horizontal strain at the bottom of the asphalt layer, the vertical compressive strain on top of the soil subgrade, and the vertical compressive strain on top of the EPS block. The Shell Pavement Design Manual (22) provides maximum allowable strain values based on the allowable number of load applications. After the strains of the various layers are calculated, the pavement design life can be determined in terms of the allowable number of 100 kN (22.5 kip) axle load repetitions (10). As part of the research reported herein, pavement design catalogs were developed to facilitate pavement system design. A design catalog is a means for designers to obtain pavement structural designs based on a unique set of assumptions relative to design requirements (23). The AASHTO 1993 design procedure was used to develop a flexible and rigid pavement design catalog. The AASHTO design procedure was used because most pavement designers in the U.S. are familiar with this design procedure. However, as with any pavement design, the limitations of the design procedure utilized need to be considered. This is especially true when designing a pavement system over EPS because traditional pavement design procedures were developed for pavement design over a soil subgrade. Therefore, the design obtained using one procedure should be checked by using other design methods, modeling tools, and analytical tools. As more

4-20 performance data of pavements over geofoam in the U.S. becomes available, the typical design methods used in the U.S. can be refined and adapted to an EPS subgrade. For pavement design procedures that do not require explicit calculations of stresses and strains within the various layers of the pavement system, it is recommended that the vertical stress on top of the EPS blocks be estimated to verify that it does not exceed the elastic limit stress of the EPS blocks. This checking can be performed utilizing a procedure similar to the one suggested for load bearing analysis of the EPS blocks outlined in Chapter 6. FLEXIBLE PAVEMENT SYSTEM DESIGN CATALOG A design catalog or design chart was developed herein for flexible pavements. A design catalog is a means for designers to obtain pavement structural designs based on a unique set of assumptions relative to design requirements (23). The flexible pavement design catalog, which is shown in Table 4.2 and developed herein, is based on the following assumptions: 1. All designs are based on the structural requirement for one performance period, regardless of the time interval. Performance period is defined as the period of time that an initial (or rehabilitated) structure will last before reaching its terminal serviceability (23). 2. The range of traffic levels for the performance period is limited to between 50,000 and 1,000,000 80 kN (18 kip) ESAL (equivalent single axle loads) applications. An ESAL is the summation of equivalent 80 kN (18 kip) single axle loads used to combine mixed traffic to design traffic for the performance period (23). 3. The designs are based on either a 50- or-75-percent level of reliability, which AASHTO considers acceptable for low-volume road design. 4. The designs are based on the resilient modulus values indicated in Table 4.1 for the three typical grades of EPS: EPS50, EPS70, and EPS100. 5. The designs are based on an initial serviceability index of 4.2 and a terminal serviceability index of 2. The average initial serviceability at the AASHO (American Association of State

4-21 Highway Officials) road test was 4.2 for flexible pavements. AASHTO recommends a terminal serviceability index of 2 for highways with lesser traffic than major highways. 6. The designs are based on a standard deviation of 0.49 to account for variability associated with material properties, traffic, and performance. AASHTO recommends a value of 0.49 for the case where the variance of projected future traffic is not considered. 7. The designs do not consider the effects of drainage levels on predicted pavement performance. Table 4.2 is similar in format to the design catalogs provided in (23). Although the design catalog in Table 4.2 is for low-volume roads, the use of EPS-block geofoam is not limited to low- volume roads and has been used for high-volume traffic roads such as interstate highways. For example, in the U.S.A., EPS-block geofoam has been used for portions of the Interstate 15 project in Salt Lake City, Utah. A design catalog is typically provided for low-volume roads to limit the number of design variables required in developing pavement structural designs. Table 4.2. Flexible Pavement Design Catalog for Low-Volume Roads. Once a design structural number (SN) is determined, appropriate flexible pavement layer thicknesses can be identified that will yield the required load-carrying capacity indicated by the structural number in accordance with the following AASHTO flexible pavement design equation: SN = a1D1 + a2D2 + a3D3 (4.1) where a1, a2, and a3 = layer coefficients for surface, base, and subbase course materials, respectively, and D1, D2, and D3 = thickness (in inches) of surface, base, and subbase course, respectively. Layer coefficients can be obtained in (23) or from state department of transportation design manuals. However, layer coefficient values for PCC slabs are not provided in (23). If a reinforced PCC slab is considered as a separation layer between the top of the EPS blocks and the overlying pavement system, it may be possible to incorporate the PCC slab into the AASHTO

4-22 1993 flexible pavement design procedure by determining a suitable layer coefficient to represent the PCC slab. In (24), it is indicated that based on test results performed in Illinois, a PCC base with a 7-day strength of 17.2 MPa (2,500 lbs/in²) exhibits a layer coefficient of 0.5. It can be seen that for a given set of layer coefficients, Equation (4.1) does not provide a unique solution of the surface, base, and subbase thicknesses. Cost effectiveness, construction, and maintenance constraints must be considered to produce a practical design (23). However, some guidance is available for estimating these thicknesses from AASHTO. For example, AASHTO recommends the minimum thickness values indicated in Table 4.3 for asphalt concrete and aggregate base to overcome placement impracticalities, ensure adequate performance, and for economic reasons. Additionally, these recommended minimum thicknesses consider the minimum layer thickness requirements for stability and cohesion under traffic loadings. This provides guidance in fixing a value of D1 and D2 so D3 can be estimated in Equation (4.1). In addition, it is recommended herein that a minimum pavement system thickness of 610 mm (24 in.) be used over EPS-block geofoam to minimize the potential for differential icing and solar heating. After various layer thickness combinations have been determined and checked against construction and maintenance constraints, a cost-effective layer thickness combination is typically selected. Table 4.3. Minimum Practical Thicknesses for Asphalt Concrete and Aggregate Base (23). A sensitivity analysis was performed to demonstrate the effects of varying EPS types at the top of the fill mass on the design structural number (SN). The input values for the sensitivity analysis are shown in Table 4.4. The sensitivity analysis was performed by varying the resilient modulus of the EPS while keeping the remainder of the input variables constant. Although the resilient modulus for the highest density EPS type, EPS100, is approximately 10 MPa (1,450 lbs/in²), higher resilient modulus values were also considered in the sensitivity analysis to analyze any potential behavioral trend.

4-23 Table 4.4. Standard Set of Inputs for Sensitivity Analysis. Figure 4.1 illustrates the effect of the EPS block resilient modulus on the design SN. The effect on the design SN is greater at lower resilient modulus values within the range of typical EPS types. The effect of varying the resilient modulus decreases at higher modulus values especially at resilient modulus values greater than 13.8 MPa (2,000 lbs/in²). From Figure 4.1, the required SN is 7.5, 6.8, and 6.1 for an EPS50, 70, and 100, respectively. Thus, the change in SN between EPS types is 0.7. Therefore, a decrease in structural number may be obtained by using an EPS with a higher density. The sensitivity results with respect to changes in resilient modulus is of special interest because these results indicate the effects of geofoam density on the structural number which can then be used to reduce the cost of the pavement system. This can be accomplished by considering the cost impact of using a higher density geofoam as the upper layer of the fill mass or increasing the thickness of the pavement structure which would impact internal and external stability. Figure 4.1. Sensitivity of AASHTO design procedure to resilient modulus. In order to further investigate the potential technical and cost benefits of utilizing an EPS with a higher density for the uppermost layer of blocks within the horizontal limits of the paved area, an example pavement system design was performed based on the design inputs shown in Table 4.4. The pavement system was assumed to consist of asphalt concrete and a crushed stone base. Based on an assumed asphalt concrete thickness of 102 mm (4 in.) with a layer coefficient of 0.44 and a crushed stone base with a layer coefficient of 0.14, the thickness of crushed stone base required would be 1,041 mm (41 in.), 914 m (36 in.), and 787 mm (31 in.) if EPS50, 70, and 100, respectively, was used for the top layer in the embankment directly underlying the pavement system. Thus, as shown in Figure 4.2, each increase in EPS grade translates into about a 127 mm (5 in.) decrease of crushed stone base. Figure 4.2. Effect of EPS density on the required base thickness for design example

4-24 for an asphalt concrete thickness of 102 mm (4 in.). Table 4.5 provides a summary of a cost comparison for this example. The use of EPS50 block is the baseline for comparison against the other two EPS types. Using an EPS70 would decrease the required crushed stone thickness by 127 mm (5 in.) and would cost about $6.71 per m2 ($5.62 per yd2) more than the EPS50 alternative. However, this cost comparison does not include the cost associated with the time savings of placing EPS blocks versus placing and compacting crushed stone. Using an EPS100 would decrease the required crushed stone thickness by 254 mm (10 in.) and would cost about $19.15 per m2 ($16.01 per yd2) more than the EPS50 alternative. The final decision as to which EPS type to select will be based on the impact the dead load stresses of each alternative will have on external and internal stability. Assuming that all three alternatives satisfy external and internal stability requirements EPS50 may be the most economical if placement and compaction of the crushed stone base is not considered. This example suggests that the use of a higher density EPS type for the uppermost layer of blocks may not be cost beneficial for low-volume roads but may be cost beneficial for high-volume roads. Table 4.5 . Example cost comparison between using various EPS grades versus using additional crushed stone base. RIGID PAVEMENT SYSTEM DESIGN CATALOG Design catalogs for rigid pavements developed herein and based on the AASHTO 1993 design procedure are presented in Tables 4.6 and 4.7. The rigid pavement design catalogs are similar to the rigid pavement design catalogs provided in (23) except that the designs are based on the resilient modulus values representative of an EPS subgrade shown in Table 4.1. Tables 4.6 and 4.7 can be used by design engineers to obtain a concrete thickness with a geofoam embankment. As with the design catalogs provided in (23), Tables 4.6 and 4.7 are based on the following assumptions: • Slab thickness design recommendations apply to all six U.S. climatic regions.

4-25 • The procedure is based on the use of dowels at transverse joints. • The range of traffic loads for the performance period is limited to between 50,000 and 1,000,000 applications of 80 kN (18 kip) ESALs (equivalent single axle loads). An ESAL is the summation of equivalent 80 kN (18 kip) single axle loads used to combine mixed traffic to design traffic for the performance period (23). • The designs are based on either a 50-percent and 75-percent level of reliability, which AASHTO considers acceptable for low-volume road design. • The designs are based on a minimum thickness of high quality material subbase equivalent to 610 mm (24 in.) less the PCC slab thickness used. This provides a minimum recommended pavement system thickness over the EPS blocks of 610 mm (24 in.) to minimize the potential for differential icing and solar heating. • The designs are based on the resilient modulus values indicated in Table 4.1 for EPS70 and 100. EPS40 is not recommended directly beneath paved areas. EPS50 was not considered because the design chart for estimating the composite modulus of subgrade reaction included in (23) does not consider a roadbed soil resilient modulus of less than 6.9 MPa (1,000 lbs/in2). • The designs are based on a mean PCC modulus of rupture (S’c) of 4.1 or 4.8 MPa (600 or 700 lbs/in2). • The designs are based on a mean PCC elastic modulus (Ec) of 34.5 GPa (5,000,000 lbs/in2). • Drainage (moisture) conditions are fair (Cd = 1.0). • The 80 kN (18-kip) ESAL traffic levels are: • High 700,000 to 1,000,000 • Medium 400,000 to 600,000

4-26 • Low 50,000 to 300,000 • A factor termed the “loss of support” factor is included in the AASHTO rigid pavement design procedure to account for the potential loss of support resulting from base and subbase erosion and/or differential vertical soil movements. Loss of support factors for typical base and subbase materials range between 0 and 3. A loss of support factor of 0 indicates that no loss of pavement support is anticipated and is indicated in the design catalog as edge support equal to “yes”. A value greater than 0 indicates that some loss of support may occur. In the design catalog, a loss of support is indicated as edge support equal to “no” and the design is based on a loss of support factor of 3. Tables 4.6 and 4.7 are similar in format to the design catalogs provided in (23). As discussed previously, for the flexible pavement design catalog, although the design catalog in Tables 4.6 and 4.7 are for low-volume roads, EPS-block geofoam can be and has been used for high-volume traffic roads such as interstate highways. A design catalog is typically provided for low-volume roads to limit the number of design variables required in developing pavement structural designs. Table 4.6. Rigid pavement design catalog for low-volume roads for inherent reliability of 50 percent. Table 4.7. Rigid pavement design catalog for low-volume roads for inherent reliability of 75 percent. SUMMARY OF PAVEMENT SYSTEM DATA FROM CASE HISTORIES Table 4.8 provides a summary of various pavement system designs that have been utilized in the U.S.A. over EPS-block geofoam. Both embankment and bridge approach case histories are included in Table 4.8. These case histories are discussed in Chapter 11. Based on the five case

4-27 histories for which pavement thickness data is available, the total pavement system thicknesses range from 508 to 864 mm (20 to 34 in.) and average 660 mm (26 in.). Table 4.8. Summary of EPS-block geofoam pavement system data from case histories. TYPICAL DEAD LOAD STRESS RANGE IMPOSED BY A PAVEMENT SYSTEM The proposed EPS-block geofoam embankment design procedure discussed in Chapter 3 requires that a preliminary pavement system design be assumed to estimate the gravity loads for use in the external and internal stability analyses prior to performing the final pavement design. Therefore, it is useful to establish a dead load stress range that a typical pavement system may impose on the EPS-blocks for use in preliminary internal and external stability analysis. Two approaches were used to investigate the dead load stresses imposed by typical pavement systems on an EPS-block geofoam embankment. The first approach was to analyze pavement system designs based on the AASHTO 1993 flexible and rigid pavement design procedures for low-volume roads. The second approach was to investigate pavement system data from EPS-block geofoam case histories. Flexible pavement systems were designed herein by assuming the pavement system over the EPS blocks consists of an asphalt concrete surface, a crushed stone base, and a sandy gravel subbase. The AASHTO minimum recommended thicknesses for asphalt concrete and aggregate base, shown in Table 4.3, were used to determine the thickness of subbase required to provide the required SN. Table 4.9 provides the pavement material layer coefficients and compacted unit weight values that were assumed. The unit weight of the asphalt concrete represents a bulk unit weight and not the maximum theoretical unit weight, the later being the unit weight value that would be obtained if the bituminous layer was compacted such that no voids would remain in the aggregate-bitumen mixture. The unit weight values for the crushed stone base and the sandy gravel subbase were estimated from typical values of optimum moisture content and maximum

4-28 dry unit weights based on the ASTM D 698 laboratory procedure indicated in (26) and a compaction effort of 97 percent of the maximum dry unit weight. For each traffic range shown in Table 4.3, the subbase thickness required for the SN values of 3 to 9 was determined. The minimum SN value was selected based on the minimum value of 3.1 obtained for the flexible pavement design catalog for low-volume roads with an EPS subgrade (Table 4.2) The maximum value of 9 was selected because this is the maximum value provided in the AASHTO nomograph provided in (23) for obtaining the design SN for a flexible pavement. Table 4.9. Material layer coefficients and density values used in the analyses of flexible pavements. For SN values ranging from 3 to 9, an asphalt concrete pavement, and the base thicknesses in Table 4.3, the overall pavement system thickness was less than the minimum pavement system thickness of 610 mm (24 in.), which is recommended over EPS-block geofoam embankments to minimize the potential for differential icing and solar heating. In order to obtain the recommended 610 mm (24 in.) pavement system thickness, the thickness of the sandy gravel subbase was increased to a thickness that would provide a pavement section thickness of 610 mm (24 in.). Table 4.10 presents a summary of the dead load stresses and average unit weight values of the final flexible pavement systems determined. Stresses ranged from 12.6 to 12.9 kPa (263 to 269 lbs/ft2) and the average unit weight values ranged from 20.8 to 21.2 kN/m3 (132 to 135 lbf/ft3) for the case of a flexible pavement system without a PCC separation layer. The additional stress imposed by a PCC slab separation layer was also investigated. However, the PCC slab was not considered in the design of the pavement system, i.e., the strength contribution of the PCC slab was not considered. The additional stress from the PCC slab was determined by replacing a thickness of subbase equivalent to the thickness of the PCC slab. Thus, in all cases the minimum

4-29 recommended pavement system thickness of 610 mm (24 in.) was maintained. Table 4.10 also shows the resulting overall stresses and average unit weights for a flexible pavement system with a 102 and 152 mm (4 and 6 in.) PCC separation layer. Based on an assumed unit weight for the PCC slab of 23.6 kN/m3 (150 lbf/ft3), additional stresses imposed by the PCC slab of 0.3 and 0.5 kPa (6.7 and 10 lbs/ft2) for a 102 and 152 mm (4 and 6 in.) slab, respectively, were calculated. These supplemental stresses yielded average unit weight values for the entire pavement system ranging between 21.3 and 22.0 kN/m3 (135 and 140 lbf/ft3). Table 4.10. Typical dead load stress range for a low-volume road flexible pavement system. Rigid pavement systems were designed by assuming the PCC slab thicknesses to be between 127 and 216 mm (5 and 8.5 in.) because these PCC slab thicknesses were obtained in the rigid pavement design catalogs (Tables 4.6 and 4.7). A total pavement system thickness of 610 mm (24 in.) was used in all cases because this is the minimum recommended pavement system thickness to minimize the potential for differential icing and solar heating. The thickness of the crushed stone base was determined by taking the difference between 610 mm (24 in.) and the PCC slab thickness. A material unit weight value of 23.6 and 21.7 kN/m3 (150 and 138 lbf/ft3) was assumed for the PCC slab and the subbase, respectively. Table 4.11 presents a summary of the dead load stresses and average unit weight values of typical rigid pavement systems. As shown in Table 4.11, stresses obtained ranged from 13.5 to 13.6 kPa (282 to 284 lbs/ft2) and the average unit weight values ranged from 22.1 to 22.3 kN/m3 (140 to 142 lbf/ft3). Table 4.11. Typical dead load stress range for a low-volume road rigid pavement system. The results of the analysis for flexible pavement design for low-volume roads, revealed that the thickness of the pavement system was controlled by the minimum thickness of 610 mm (24 in.) recommended to minimize the potential for differential icing and solar heating. The analysis of both flexible and rigid pavement systems indicate that overall average pavement system material unit weight values ranging from 20.8 to 22.3 kN/m3 (132 to 142 lbf/ft3). These

4-30 average unit weights are based on a 610 mm (24 in.) overall pavement system thickness and low- volume traffic, which is defined in (23) as less than 1,000,000 applications of 80 kN (18-kip) ESALs. Average dead load stresses calculated ranged from 12.6 to 13.6 kPa (263 to 284 lbs/ft2). The case history data shown in Table 4.8 shows total pavement system thicknesses ranging from 508 to 864 mm (20 to 34 in.). This is an overall average of 660 mm (26 in.). Case history data of EPS-block geofoam projects in Norway reported in (27) indicate an average pavement system thickness of 660 mm (26 in.). The proposed design procedure outlined in Figure 3.3 is based on obtaining a pavement system that provides the least amount of stress on top of the EPS-block geofoam embankment to satisfy internal and external stability requirements. Therefore, it is recommended that the various component layers of the pavement system be assumed to have a total (wet) unit weight of 20 kN/m3 (130 lbs/ft3) for initial design purposes, which is the approximate unit weight of typical less costly subbase materials. It is also recommended that the pavement system be assumed to have a thickness of 610 mm (24 in.) for preliminary external and internal stability analyses. Alternatively, the design catalogs (Tables 4.2, 4.6, and 4.7) can be used to obtain a pavement system that can be used for the external and internal stability analyses. SUMMARY The thickness of the pavement system will affect both external and internal stability of the embankment. The benefits of using a thicker pavement system include increased life of the pavement and factor of safety of certain failure mechanisms affecting internal stability of the embankment (see Table 3.1), reduced potential for differential icing and solar heating, and better accommodation of shallow utilities and road hardware. However, a thicker pavement system will yield a heavier pavement system. This increase in weight will decrease the factor of safety of certain external stability failure modes of the embankment (see Table 3.1). Thus, some compromise is required to optimize the final design of both the pavement system and overall fill.

4-31 Regardless of the design process used, the goal of the pavement design should be to use the most economical arrangement and thickness of each material to protect the pavement from distress caused by both traffic loads and the environment (1). However, a unique aspect of pavement design over lightweight fill is that the design must also consider the external and internal stability of the embankment. As indicated in Figure 3.3, a preliminary pavement system must be assumed to perform external and internal stability analysis. Gravity loads can be calculated based on a preliminary assumed cross-section of the embankment, including the pavement system, and any cover over the sides of the embankment. Although the pavement system has not been designed at this point, it should be greater than 610 mm (24 in.) in thickness to minimize the effects of differential icing and solar heating. The design procedure depicted in Figure 3.3 is based on obtaining a pavement system that provides the least amount of stress on top of the EPS-block geofoam embankment to satisfy external and internal stability requirements. Therefore, it is recommended that the preliminary pavement system be assumed to be 610 mm (24 in.) thick and the various component layers of the pavement system be assumed to have a total (moist) unit weight of 20 kN/m3 (130 lbf/ft3) for initial design purposes. Alternatively, the design catalogs (Tables 4.2, 4.6, and 4.7) can be used to obtain a pavement system that can be used for external and internal stability analysis. Results from preliminary analysis performed during this study indicate that the use of a higher density EPS for the uppermost layer of blocks may not be cost beneficial for low-volume roads but may be cost beneficial for high-volume roads. REFERENCES 1. ERES Consultants Inc., “Pavement Design, Principles and Practices, A Training Course Participant Notebook.” Federal Highway Administration (1987) . 2. Beer, F. P., and Johnston Jr., R., Mechanics of Materials, , McGraw-Hill, Inc., New York (1981) 616 pp. 3. West, H. H., Analysis of Structures, , John Wiley & Sons, Inc., New York (1980) 689 pp. 4. Horvath, J. S., “Non-Earth Subgrade Materials and Their Thermal Effects on Pavements: An Overview.” Report No. CGT-2001-2, Manhattan College, Bronx, NY (2001) .

4-32 5. Horvath, J. S., Geofoam Geosynthetic, Horvath Engineering, P.C., Scarsdale, NY (1995) 229 pp. 6. Refsdal, G., “Frost Protection of Road Pavements.” Frost Action in Soils - No. 26, Committee on Permafrost, ed., Oslo, Norway (1987) pp. 3-19. 7. Gandahl, R., “Polystyrene Foam as a Frost Protection Measure on National Roads in Sweden.” Transportation Research Record No. 1146, Transportation Research Board, Washington, D.C. (1987) pp. 1-9. 8. “Expanded Polystyrene Used in Road Embankments - Design, Construction and Quality Assurance.” Form 482E, Public Roads Administration, Road Research Laboratory, Oslo, Norway (1992) 4 pp. 9. Duskov, M., “EPS as a Light-Weight Sub-Base Material in Pavement Structures,” Doctor of Engineering thesis, Delft University of Technology, Delft, The Netherlands (1998). 10. Duskov, M., “Dutch Design Manual for Lightweight Pavements with Expanded Polystyrene Geofoam.” Transportation Research Record 1736, Tranportation Research Board, Washington, D.C. (2000) pp. 103-109. 11. Horvath, J. S., “Is a Concrete Slab Really Necessary for EPS-Geofoam Lightweight Fills: Myth versus Reality,” In Manhattan College-School of Engineering, Center for Geotechnolgy [website]. [updated 9 August 2001; cited 20 September2001]. Available from http://www.engineering.manhattan.edu/civil/CGT/T2olrgeomat3.html; INTERNET. 12. “Merkblatt für die Verwendung von EPS-Hartschaumstoffen beim Bau von Straßendämmen.” Forschungsgesellschaft für Straßen- und Verkehrswesen, Arbeitsgruppe Erd- und Grundbau, Köln, Deutschland (1995) 27 pp. 13. “Code of Practice; Using Expanded Polystyrene for the Construction of Road Embankments.” BASF AG, Ludwigshafen, Germany (1995) 14 pp. 14. Sanders, R. L., and Seedhouse, R. L., “The Use of Polystyrene for Embankment Construction.” Contractor Report 356, Transport Research Laboratory, Crowthorne, Berkshire, U.K. (1994) 55 pp. 15. Newmark, N. M., “Effects of Earthquakes on Dams and Embankments.” Geotechnique, Vol. 15, No. 2 (1965) pp. 139-159. 16. Bull-Wasser, R., “EPS-Hartschaum als Baustoff für Straßen.” Straßenbau Heft S 4, Berichte der Bundesanstalt für Straßenwesen, Bergisch Gladbach, Germany (1993) 156 pp. 17. Duskov, M., “Falling Weight Deflection Measurements on Asphalt Test Pavements with EPS at the Bundesanstalt für Straßenwesen.” EPS-Hartschaum als Baustoff für Straßen, R. Bull-Wasser, ed. Berichte der Bundesanstalt für Straßenwesen, Bergisch Gladbach, Germany (1993) pp. 129-156. 18. Duskov, M., and Bull-Wasser, R., “Analysis of Asphalt Test Pavements with a Sub-Base of Expanded Polystyrene Foam.” Proceedings of the 7th International Conference on Asphalt Pavements-Design, Construction and Performance,1992, Nottingham, Vol. III (1992) pp. 96-109. 19. Duskov, M., “Materials Research on Expanded Polystyrene Foam (EPS).” Report 7-94- 211-2, Delft University of Technology, Delft, The Netherlands (1993) . 20. Duskov, M., “DIANA Non-linear Analysis of Pavement Structures with an EPS Sub-base under Static Loading.” Report 7-94-211-3, Delft University of Technology, Delft, The Netherlands (1994) 41 pp. 21. Huang, Y. H., “Pavement Analysis and Design.” Prentice-Hall, Inc., Englewood Cliffs, NJ, (1993) 805. 22. “Shell Pavement Design Manual.” Shell International Petroleum Company Limited, London (1983) .

4-33 23. American Association of State Highway and Transportation Officials, AASHTO Guide for Design of Pavement Structures, 1993, , American Association of State Highway and Transportation Officials, Washington, D.C. (1993) . 24. Van Til, C. J., McCullough, B. F., Vallerga, B. A., and Hicks, R. G., “Evaluation of AASHO Interim Guides for Design of Pavement Structures.” NCHRP Report 128, Transportation Research Board, Washington, D.C. (1972) . 25. RSMeans Company Inc., RSMeans Site Work & Landscape Cost Data, 20th Annual Edition, 2001, , RSMeans Company Inc., Kingston, MA (2000) 638 pp. 26. Lindeburg, M. R., Civil Engineering Reference Manual, 4th, Professional Publications, Inc., San Carlos, CA (1986) . 27. Aabøe, R., “13 years of experience with EPS as a lightweight fill material in road embankments.” Plastic Foam in Road Embankments, Publication No. 61, Norwegian Road Research Laboratary, Oslo, Norway (1987) pp. 21-27.

FIGURE 4.1 PROJ 24-11.doc Resilient Modulus (MPa) 0 5 10 15 20 25 30 St ru ct u ra l N u m be r, SN 0 1 2 3 4 5 6 7 8 4-34

FIGURE 4.2 PROJ 24-11.doc EPS Density (kg/m3) 18 20 22 24 26 28 30 32 34 Cr u sh ed S to n e B as e Th ic kn es s (m m) 750 800 850 900 950 1000 1050 1100 4-35

TABLE 4.1 PROJ 24-11.doc Design Values of Engineering Parameters Proposed AASHTO Material Designation Minimum Allowable Full Block Density, kg/m3(lbf/ft3) CBR (%) Initial Tangent Young's Modulus, Eti MPa(lbs/in2) Resilient Modulus, MR MPa(lbs/in2) EPS50 20 (1.25) 2 5 (725) 5 (725) EPS70 24 (1.5) 3 7 (1015) 7 (1015) EPS100 32 (2.0) 4 10 (1450) 10 (1450) Note: The use of EPS40 directly beneath paved areas is not recommended and thus does not appear in this table because of the potential for settlement problems. The minimum allowable block density is based on density obtained on a block as whole or full-sized block. The proposed AASHTO material type designation system, which is presented in Chapter 9, is based on the minimum elastic limit stress of the block as a whole in kilopascals. 4-36

TABLE 4.2 PROJ 24-11.doc R EPS Type Traffic Level (%) Low Medium High 50,000 300,000 400,000 600,000 700,000 1,000,000 50 EPS50 4* 5.1 5.3 5.5 5.7 5.9 EPS70 3.5 4.6 4.7 5 5.1 5.3 EPS100 3.1 4.1 4.2 4.5 4.6 4.8 75 EPS50 4.4 5.6 5.8 6.1 6.2 6.5 EPS70 3.9 5 5.2 5.5 5.6 5.9 EPS100 3.5 4.5 4.7 5 5.1 5.3 Note: R = Reliability level. * design structural number, SN. 4-37

TABLE 4.3 PROJ 24-11.doc Minimum Thickness , mm (in.) Traffic, ESALs Asphalt Concrete Aggregate Base Less than 50,000 25 (1.0) 100 (4.0) 50,001 – 150,000 50 (2.0) 100 (4.0) 150,001 – 500,000 64 (2.5) 100 (4.0) 500,001 – 2,000,000 76 (3.0) 150 (6.0) 2,000,001 – 7,000,000 90 (3.5) 150 (6.0) Greater than 7,000,000 100 (4.0) 150 (6.0) 4-38

TABLE 4.4 PROJ 24-11.doc Variable Initial or Constant Input Value (18-kip) ESALs Over Initial Performance Period 1,000,000 Initial Serviceability 4.2 Terminal Serviceability 2.5 Reliability Level (%) 90 Overall Standard Deviation 0.44 EPS Block Resilient Modulus Varied 4-39

TABLE 4.5 PROJ 24-11.doc EPS Type Total Crushed Stone Thickness mm (in.) EPS Cost $/m2 ($/yd2) EPS Cost Difference between EPS types $/m2 ($/yd2) Crushed Stone Cost $/m2 ($/yd2) Crushed Stone Cost Difference between EPS types $/m2 ($/yd2) Cost Difference Between Higher EPS Grade and Crushed Stone $/m2 ($/yd2) EPS50 1,041 (41) 26.21 (22.42) - 27.08 (22.64) - - EPS70 914 (36) 30.48 (25.49) 4.27 (3.07) 23.77 (19.87) -3.31 (-2.77) 6.71 (5.62) EPS100 787 (31) 39.62 (33.13) 13.41 (10.71) 20.47 (17.12) -6.61 (-5.52) 19.15 (16.01) Note: Cost of EPS is based on a cost of $43.00, $50.00, and $65.00 per cubic meter ($32.88, $38.23 $49.70 per cubic yard) for EPS50, 70, and 100, respectively, and on a 0.61m (24 in.) block thickness. See Chapter 12 for EPS cost information. Cost of crushed stone is based on a cost of $26.00 per cubic meter ($19.88 per cubic yard) obtained from (25). 4-40

TABLE 4.6 PROJ 24-11.doc LOAD TRANSFER DEVICES No Yes EDGE SUPPORT No Yes No Yes S'c MPa (lbs/in2) 4.1 (600) 4.8 (700) 4.1 (600) 4.8 (700) 4.1 (600) 4.8 (700) 4.1 (600) 4.8 (700) INHERENT RELIABILITY EPS TYPE EPS RESILIENT MODULUS Traffic % MPa (lbs/in2) ESALs 50 EPS70 7 (1015) 50,000 5 5 5 5 5 5 5 5 EPS100 10 (1450) 5 5 5 5 5 5 5 5 EPS70 7 (1015) 300,000 6.5 6 6 6 6 5.5 5.5 5 EPS100 10 (1450) 6.5 6 6 6 6 5.5 5.5 5 EPS70 7 (1015) 400,000 7 6.5 6.5 6 6 5.5 6 5.5 EPS100 10 (1450) 7 6.5 6.5 6 6 5.5 6 5.5 EPS70 7 (1015) 600,000 7.5 7 7 6.5 6.5 6 6 5.5 EPS100 10 (1450) 7.5 7 7 6.5 6.5 6 6 5.5 EPS70 7 (1015) 700,000 7.5 7 7 6.5 6.5 6 6 6 EPS100 10 (1450) 7.5 7 7 6.5 6.5 6 6 6 EPS70 7 (1015) 1,000,000 8 7.5 7.5 7 7 6.5 6.5 6 EPS100 10 (1450) 8 7.5 7.5 7 7 6.5 6.5 6 4-41

TABLE 4.7 PROJ 24-11.doc LOAD TRANSFER DEVICES No Yes EDGE SUPPORT No Yes No Yes S'c MPa (lbs/in2) 4.1 (600) 4.8 (700) 4.1 (600) 4.8 (700) 4.1 (600) 4.8 (700) 4.1 (600) 4.8 (700) INHERENT RELIABILITY EPS TYPE EPS RESILIENT MODULUS Traffic % MPa (lbs/in2) ESALs 75 EPS70 7 (1015) 50,000 5.5 5 5 5 5 5 5 5 EPS100 10 (1450) 5.5 5 5 5 5 5 5 5 EPS70 7 (1015) 300,000 7 6.5 6.5 6 6.5 6 6 5.5 EPS100 10 (1450) 7 6.5 7 6 6.5 6 6 5.5 EPS70 7 (1015) 400,000 7.5 7 7 6.5 6.5 6 6 6 EPS100 10 (1450) 7.5 7 7 6.5 6.5 6 6 6 EPS70 7 (1015) 600,000 8 7.5 7.5 7 7 6.5 6.5 6 EPS100 10 (1450) 8 7.5 7.5 7 7 6.5 6.5 6 EPS70 7 (1015) 700,000 8 7.5 7.5 7 7 6.5 7 6 EPS100 10 (1450) 8 7.5 7.5 7 7 6.5 7 6 EPS70 7 (1015) 1,000,000 8.5 8 8 7.5 7.5 7 7 6.5 EPS100 10 (1450) 8.5 8 8 7.5 7.5 7 7 6.5 4-42

TABLE 4.8 PROJ 24-11.doc New York: State Route 23A, Town of Jewett, Greene County: 230 mm (9 in.) asphalt pavement 381 mm (15 in.) graded crushed-stone subbase 100 mm (4 in.) reinforced-concrete cap Total Pavement System Thickness=711 mm (28 in.) 2.8 m (9 ft) of 20 kg/m3 (1.25 lbf/ft3) EPS Utah: I-15 Reconstruction Pavement layer thicknesses varied. PCCP Open graded base Dense graded base Granular borrow Load distribution slab designed for HS 20 loading Total Pavement System Thickness=varies Various thicknesses of 18 kg/m3 minimum (1.12 lbf/ft3) EPS Illinois: 143rd Street, Orland Park: 44 mm (1.75 in.) bituminous concrete surface 38 mm (1.5 in.) bituminous concrete binder 229 mm (9 in.) PCC Base 165 mm (6.5 in.) over crown Aggregate Subgrade but varies to accommodate a crowned roadway. 102 mm (4 in.) PCC Base Special with welded wire fabric Total Pavement System Thickness=578 mm (22.75 in.) 0.9 - 1.2 m (3 - 4 ft.) of 24 kg/m3 (1.5 lbf/ft3) EPS Wyoming: Bridge Rehabilitation, N.F. Shoshone River 50 mm (2 in.) plant mix bit. 255 mm (10 in.) approach reinforced PCC slab 205 mm (8 in.) min. sand Total Pavement System Thickness=510 mm (20 in.) 2.75m (9 ft) of 24kg/m3 (1.5 lbf/ft3) EPS Indiana: State Route 109, Noble County 330 mm (13 in.) bituminous pavement 406 mm (16 in.) #8 Stone 102 - 127 mm (4 in.-5 in.) reinforced concrete slab Total Pavement System Thickness=863 mm (34 in.) 0.4 - 1.5 m (1.25 - 5 ft) of 24 kg/m3 (1.5 lbf/ft3) EPS Wyoming: Moorcraft Bridge, Crook County 305 mm (12 in.) Asphalt pavement and Concrete approach slab 305 mm (12 in.) Crushed base Impermeable membrane Total Pavement System Thickness=610 mm (24 in.) 1.2 m (4 ft) 24 kg/m3 (1.5 lbf/ft3) EPS Note: PCC is Portland Cement Concrete and PCCP is Portland Cement Concrete Pavement. 4-43

TABLE 4.9 PROJ 24-11.doc Material Layer Coefficient (1) Unit Weight kN/m3 (lbf/ft3) Surface Course or Base Asphalt Concrete- plant mix 0.44 23.3 (148) Base Crushed Stone 0.14 21.7 (138) Subbase Sandy gravel 0.11 20.4 (130) (1) Layer coefficients obtained from (26). 4-44

TABLE 4.10 PROJ 24-11.doc Crushed Sandy No PCC Slab Separation Layer 4-In PCC Slab Separation Layer 6-In PCC Slab Separation Layer Asphalt Stone Gravel Avg. Avg. Avg. Concrete Base Subbase Stress Unit Weight Stress Unit Weight Stress Unit Weight mm (in.) mm (in.) mm (in.) kPa (lbs/ft2) Kg/m3 (lbf/ft3) kPa (lbs/ft2) kg/m3 (lbf/ft3) kPa (lbs/ft2) kg/m3 (lbf/ft3) 25.4 (1) 101.6 (4) 482.6 (19) 12.6 (263) 20.8 (132) 12.9 (269) 21.3 (135) 13.1 (274) 21.5 (136) 50.8 (2) 101.6 (4) 457.2 (18) 12.7 (265) 20.9 (133) 13 (272) 21.4 (136) 13.2 (276) 21.7 (138) 63.5 (2.5) 152.4 (6) 444.5 (17.5) 12.8 (267) 20.9 (133) 13.1 (274) 21.5 (136) 13.3 (278) 21.7 (138) 76.2 (3) 152.4 (6) 381 (15) 12.9 (269) 21.1 (134) 13.2 (276) 21.6 (137) 13.4 (280) 21.9 (139) 88.9 (3.5) 152.4 (6) 368.3 (14.5) 12.9 (269) 21.1 (134) 13.2 (276) 21.7 (138) 13.4 (280) 21.9 (139) 101.6 (4) 152.4 (6) 355.6 (14) 12.9 (269) 21.2 (135) 13.2 (276) 21.7 (138) 13.4 (280) 22.0 (140) 4-45

TABLE 4.11 PROJ 24-11.doc PCC Crushed Slab Stone Avg. Thickness Subbase Stress Density mm (in.) Mm (in.) kPa (lbs/ft2) kg/m3 (lbf/ft3) 127 (5) 482.6 (19) 13.5 (282) 22.1 (140) 139.7 (5.5) 469.9 (18.5) 13.5 (282) 22.1 (140) 152.4 (6) 457.2 (18) 13.5 (282) 22.2 (141) 165.1 (6.5) 444.5 (17.5) 13.5 (282) 22.2 (141) 177.8 (7) 431.8 (17) 13.6 (284) 22.2 (141) 190.5 (7.5) 419.1 (16.5) 13.6 (284) 22.3 (142) 203.2 (8) 406.4 (16) 13.6 (284) 22.3 (142) 215.9 (8.5) 393.7 (15.5) 13.6 (284) 22.3 (142) 4-46